From: Christoph Hellwig <hch@lst.de>
To: Christian Zigotzky <chzigotzky@xenosoft.de>
Cc: Christoph Hellwig <hch@lst.de>,
linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-mm@kvack.org, iommu@lists.linux-foundation.org,
Olof Johansson <olof@lixom.net>
Subject: Re: Bug 205201 - overflow of DMA mask and bus mask
Date: Tue, 5 Nov 2019 17:28:56 +0100 [thread overview]
Message-ID: <20191105162856.GA15402@lst.de> (raw)
In-Reply-To: <46025f1b-db20-ac23-7dcd-10bc43bbb6ee@xenosoft.de>
On Tue, Nov 05, 2019 at 08:56:27AM +0100, Christian Zigotzky wrote:
> Hi All,
>
> We still have DMA problems with some PCI devices. Since the PowerPC updates
> 4.21-1 [1] we need to decrease the RAM to 3500MB (mem=3500M) if we want to
> work with our PCI devices. The FSL P5020 and P5040 have these problems
> currently.
>
> Error message:
>
> [ 25.654852] bttv 1000:04:05.0: overflow 0x00000000fe077000+4096 of DMA
> mask ffffffff bus mask df000000
>
> All 5.x Linux kernels can't initialize a SCSI PCI card anymore so booting
> of a Linux userland isn't possible.
>
> PLEASE check the DMA changes in the PowerPC updates 4.21-1 [1]. The kernel
> 4.20 works with all PCI devices without limitation of RAM.
Can you send me the .config and a dmesg? And in the meantime try the
patch below?
---
From 4d659b7311bd4141fdd3eeeb80fa2d7602ea01d4 Mon Sep 17 00:00:00 2001
From: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Date: Fri, 18 Oct 2019 13:00:43 +0200
Subject: dma-direct: check for overflows on 32 bit DMA addresses
As seen on the new Raspberry Pi 4 and sta2x11's DMA implementation it is
possible for a device configured with 32 bit DMA addresses and a partial
DMA mapping located at the end of the address space to overflow. It
happens when a higher physical address, not DMAable, is translated to
it's DMA counterpart.
For example the Raspberry Pi 4, configurable up to 4 GB of memory, has
an interconnect capable of addressing the lower 1 GB of physical memory
with a DMA offset of 0xc0000000. It transpires that, any attempt to
translate physical addresses higher than the first GB will result in an
overflow which dma_capable() can't detect as it only checks for
addresses bigger then the maximum allowed DMA address.
Fix this by verifying in dma_capable() if the DMA address range provided
is at any point lower than the minimum possible DMA address on the bus.
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
---
include/linux/dma-direct.h | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/include/linux/dma-direct.h b/include/linux/dma-direct.h
index adf993a3bd58..6ad9e9ea7564 100644
--- a/include/linux/dma-direct.h
+++ b/include/linux/dma-direct.h
@@ -3,6 +3,7 @@
#define _LINUX_DMA_DIRECT_H 1
#include <linux/dma-mapping.h>
+#include <linux/memblock.h> /* for min_low_pfn */
#include <linux/mem_encrypt.h>
#ifdef CONFIG_ARCH_HAS_PHYS_TO_DMA
@@ -27,6 +28,13 @@ static inline bool dma_capable(struct device *dev, dma_addr_t addr, size_t size)
if (!dev->dma_mask)
return false;
+#ifndef CONFIG_ARCH_DMA_ADDR_T_64BIT
+ /* Check if DMA address overflowed */
+ if (min(addr, addr + size - 1) <
+ __phys_to_dma(dev, (phys_addr_t)(min_low_pfn << PAGE_SHIFT)))
+ return false;
+#endif
+
return addr + size - 1 <=
min_not_zero(*dev->dma_mask, dev->bus_dma_mask);
}
--
2.20.1
next prev parent reply other threads:[~2019-11-05 16:29 UTC|newest]
Thread overview: 181+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-14 8:22 use generic DMA mapping code in powerpc V4 Christoph Hellwig
2018-11-14 8:22 ` [PATCH 01/34] powerpc: use mm zones more sensibly Christoph Hellwig
2018-12-06 14:09 ` Christoph Hellwig
2018-12-07 12:18 ` Michael Ellerman
2018-12-07 14:09 ` Christoph Hellwig
2018-12-07 15:04 ` Christian Zigotzky
2018-11-14 8:22 ` [PATCH 02/34] powerpc: allow NOT_COHERENT_CACHE for amigaone Christoph Hellwig
2018-12-06 14:09 ` Christoph Hellwig
2018-11-14 8:22 ` [PATCH 03/34] powerpc/dma: remove the unused ARCH_HAS_DMA_MMAP_COHERENT define Christoph Hellwig
2018-12-06 14:09 ` Christoph Hellwig
2018-11-14 8:22 ` [PATCH 04/34] powerpc/dma: remove the unused ISA_DMA_THRESHOLD export Christoph Hellwig
2018-12-06 14:10 ` Christoph Hellwig
2018-11-14 8:22 ` [PATCH 05/34] powerpc/dma: remove the unused dma_iommu_ops export Christoph Hellwig
2018-12-06 14:10 ` Christoph Hellwig
2018-11-14 8:22 ` [PATCH 06/34] powerpc/dma: split the two __dma_alloc_coherent implementations Christoph Hellwig
2018-12-06 14:10 ` Christoph Hellwig
2018-11-14 8:22 ` [PATCH 07/34] powerpc/dma: remove the no-op dma_nommu_unmap_{page,sg} routines Christoph Hellwig
2018-12-06 14:10 ` [PATCH 07/34] powerpc/dma: remove the no-op dma_nommu_unmap_{page, sg} routines Christoph Hellwig
2018-11-14 8:22 ` [PATCH 08/34] powerpc/dma: untangle vio_dma_mapping_ops from dma_iommu_ops Christoph Hellwig
2018-12-06 14:10 ` Christoph Hellwig
2018-11-14 8:22 ` [PATCH 09/34] powerpc/dma: handle iommu bypass in dma_iommu_ops Christoph Hellwig
2018-11-14 8:22 ` [PATCH 10/34] powerpc/pseries: unwind dma_get_required_mask_pSeriesLP a bit Christoph Hellwig
2018-11-14 8:22 ` [PATCH 11/34] powerpc/pseries: use the generic iommu bypass code Christoph Hellwig
2018-11-14 8:22 ` [PATCH 12/34] powerpc/cell: move dma direct window setup out of dma_configure Christoph Hellwig
2018-12-09 10:23 ` Michael Ellerman
2018-12-12 14:36 ` Christoph Hellwig
2018-12-14 13:29 ` Michael Ellerman
2018-12-14 16:42 ` Christoph Hellwig
2018-11-14 8:22 ` [PATCH 13/34] powerpc/cell: use the generic iommu bypass code Christoph Hellwig
2018-11-14 8:22 ` [PATCH 14/34] powerpc/dart: remove dead cleanup code in iommu_init_early_dart Christoph Hellwig
2018-12-06 14:11 ` Christoph Hellwig
2018-11-14 8:22 ` [PATCH 15/34] powerpc/dart: use the generic iommu bypass code Christoph Hellwig
2018-11-14 8:22 ` [PATCH 16/34] powerpc/powernv: remove pnv_pci_ioda_pe_single_vendor Christoph Hellwig
2018-11-14 8:22 ` [PATCH 17/34] powerpc/powernv: remove pnv_npu_dma_set_mask Christoph Hellwig
2018-11-14 8:22 ` [PATCH 18/34] powerpc/powernv: use the generic iommu bypass code Christoph Hellwig
2018-11-14 8:22 ` [PATCH 19/34] cxl: drop the dma_set_mask callback from vphb Christoph Hellwig
2018-12-06 14:10 ` Christoph Hellwig
2018-11-14 8:23 ` [PATCH 20/34] powerpc/dma: stop overriding dma_get_required_mask Christoph Hellwig
2018-11-14 8:23 ` [PATCH 21/34] powerpc/pci: remove the dma_set_mask pci_controller ops methods Christoph Hellwig
2018-11-14 8:23 ` [PATCH 22/34] powerpc/dma: remove the iommu fallback for coherent allocations Christoph Hellwig
2018-11-14 8:23 ` [PATCH 23/34] powerpc/dma: remove get_pci_dma_ops Christoph Hellwig
2018-11-14 8:23 ` [PATCH 24/34] powerpc/dma: move pci_dma_dev_setup_swiotlb to fsl_pci.c Christoph Hellwig
2018-11-14 8:23 ` [PATCH 25/34] powerpc/dma: remove max_direct_dma_addr Christoph Hellwig
2018-11-14 8:23 ` [PATCH 26/34] powerpc/dma: fix an off-by-one in dma_capable Christoph Hellwig
2018-11-14 8:23 ` [PATCH 27/34] powerpc/fsl_pci: simplify fsl_pci_dma_set_mask Christoph Hellwig
2018-11-14 8:23 ` [PATCH 28/34] dma-mapping, powerpc: simplify the arch dma_set_mask override Christoph Hellwig
2018-11-14 8:23 ` [PATCH 29/34] powerpc/dma: use phys_to_dma instead of get_dma_offset Christoph Hellwig
2018-11-14 8:23 ` [PATCH 30/34] powerpc/dma: remove dma_nommu_mmap_coherent Christoph Hellwig
2018-11-14 8:23 ` [PATCH 31/34] powerpc/dma: use generic direct and swiotlb ops Christoph Hellwig
2018-11-14 8:23 ` [PATCH 32/34] powerpc/dma: remove get_dma_offset Christoph Hellwig
2018-11-14 8:23 ` [PATCH 33/34] powerpc/dma: remove set_dma_offset Christoph Hellwig
2018-11-14 8:23 ` [PATCH 34/34] powerpc/dma: trim the fat from <asm/dma-mapping.h> Christoph Hellwig
2018-11-27 7:42 ` use generic DMA mapping code in powerpc V4 Christoph Hellwig
2018-11-27 22:05 ` Benjamin Herrenschmidt
2018-11-28 11:05 ` Michael Ellerman
2018-11-28 15:55 ` Christian Zigotzky
2018-11-28 20:02 ` Christian Zigotzky
2018-11-28 20:35 ` Michal Suchánek
2018-11-29 17:03 ` Christoph Hellwig
2019-12-19 13:54 ` Christian Zigotzky
2018-11-29 12:05 ` Christian Zigotzky
2018-11-29 15:30 ` Christian Zigotzky
2018-11-29 17:03 ` Christoph Hellwig
2018-11-29 23:07 ` Christian Zigotzky
2018-11-30 10:53 ` Christoph Hellwig
2018-11-30 12:23 ` Christian Zigotzky
2018-11-30 13:10 ` Christoph Hellwig
2018-11-30 15:29 ` Christian Zigotzky
2018-12-04 7:31 ` Christian Zigotzky
2018-12-04 9:53 ` Christian Zigotzky
2018-12-04 14:24 ` Christoph Hellwig
2018-12-05 9:44 ` Christian Zigotzky
2018-12-05 14:05 ` Christoph Hellwig
2018-12-06 10:55 ` Christian Zigotzky
2018-12-06 17:10 ` Christian Zigotzky
2018-12-06 19:36 ` Christoph Hellwig
2018-12-07 7:48 ` Christian Zigotzky
2018-12-07 13:45 ` Christian Zigotzky
2018-12-07 18:33 ` Christian Zigotzky
2018-12-08 10:29 ` Christian Zigotzky
2018-12-08 13:47 ` Christian Zigotzky
2018-12-09 14:20 ` Christian Zigotzky
2018-12-09 18:26 ` Christian Zigotzky
2018-12-10 15:54 ` Christian Zigotzky
2018-12-11 14:28 ` Christian Zigotzky
2018-12-11 18:17 ` Christian Zigotzky
2018-12-12 0:47 ` Benjamin Herrenschmidt
2018-12-12 7:03 ` Christian Zigotzky
2018-12-12 14:15 ` Christoph Hellwig
2018-12-12 14:39 ` Christian Zigotzky
2018-12-13 8:41 ` Christian Zigotzky
2018-12-13 9:10 ` Christoph Hellwig
2018-12-13 9:47 ` Christian Zigotzky
2018-12-13 11:19 ` Christian Zigotzky
2018-12-13 11:25 ` Christoph Hellwig
2018-12-13 13:34 ` Christian Zigotzky
2018-12-13 17:48 ` Christian Zigotzky
2018-12-13 21:53 ` Christian Zigotzky
2018-12-14 12:00 ` Christian Zigotzky
2018-12-14 16:45 ` Christoph Hellwig
2019-01-03 7:36 ` Christoph Hellwig
2019-01-03 19:26 ` Christian Zigotzky
2019-01-05 16:03 ` Christian Zigotzky
2019-01-09 9:31 ` Christian Zigotzky
2019-01-11 2:10 ` Christian Zigotzky
2019-01-12 18:14 ` Christian Zigotzky
2019-01-15 8:07 ` Christian Zigotzky
2019-01-15 8:49 ` Christian Zigotzky
2019-01-15 10:55 ` Christian Zigotzky
2019-01-15 13:35 ` Christoph Hellwig
2019-01-15 13:56 ` Christian Zigotzky
2019-01-15 15:17 ` Christoph Hellwig
2019-01-17 9:21 ` Christian Zigotzky
2019-01-17 9:31 ` Christoph Hellwig
2019-01-17 9:44 ` Christian Zigotzky
2019-01-18 8:35 ` Christoph Hellwig
2019-01-18 11:10 ` Christian Zigotzky
2019-01-18 11:28 ` Christoph Hellwig
2019-01-18 12:07 ` Christian Zigotzky
2019-01-18 12:18 ` Christoph Hellwig
2019-01-18 12:46 ` Christian Zigotzky
2019-01-18 12:55 ` Christoph Hellwig
2019-01-18 15:06 ` Christian Zigotzky
2019-01-19 11:40 ` Christian Zigotzky
2019-01-19 11:52 ` Christian Zigotzky
2019-01-19 13:02 ` Christoph Hellwig
2019-01-19 14:04 ` Christoph Hellwig
2019-01-21 14:38 ` Christian Zigotzky
2019-01-23 14:34 ` Christian Zigotzky
2019-01-25 13:37 ` Christian Zigotzky
2019-01-27 13:13 ` Christian Zigotzky
2019-01-28 7:04 ` Christoph Hellwig
[not found] ` <20190128070422.GA2772-jcswGhMUV9g@public.gmane.org>
2019-01-28 16:22 ` Christoph Hellwig
2019-01-28 16:22 ` Christoph Hellwig
2019-01-28 16:52 ` Christian Zigotzky
2019-01-28 16:52 ` Christian Zigotzky
2019-01-29 15:03 ` Christian Zigotzky
2019-01-29 16:14 ` Christoph Hellwig
2019-01-29 16:34 ` Christoph Hellwig
2019-01-30 4:40 ` Christian Zigotzky
2019-01-31 12:48 ` Christian Zigotzky
2019-02-01 8:04 ` Christoph Hellwig
2019-02-01 16:54 ` Christian Zigotzky
2019-02-03 16:49 ` Christian Zigotzky
2019-02-04 7:56 ` Christoph Hellwig
2019-02-04 12:13 ` Christian Zigotzky
2019-02-04 12:38 ` Christoph Hellwig
2019-02-06 13:45 ` Christian Zigotzky
2019-02-06 15:15 ` Christoph Hellwig
2019-02-06 15:16 ` Christoph Hellwig
2019-02-07 4:34 ` Christian Zigotzky
2019-02-08 9:01 ` Christian Zigotzky
2019-02-08 9:18 ` Christoph Hellwig
2019-02-08 10:59 ` Christian Zigotzky
2019-02-10 12:00 ` Christian Zigotzky
2019-02-11 7:16 ` Christian Zigotzky
2019-02-11 7:21 ` Christian Zigotzky
2019-02-11 7:38 ` Christoph Hellwig
2019-02-12 12:42 ` Christian Zigotzky
2019-02-12 15:25 ` Christoph Hellwig
2019-02-12 19:31 ` Christian Zigotzky
2019-02-12 19:50 ` Christian Zigotzky
2019-02-12 19:52 ` Christoph Hellwig
2019-11-05 7:56 ` Bug 205201 - overflow of DMA mask and bus mask Christian Zigotzky
2019-11-05 16:28 ` Christoph Hellwig [this message]
2019-11-06 14:09 ` Robin Murphy
2019-11-21 17:26 ` Christoph Hellwig
2019-11-07 9:53 ` Christian Zigotzky
2019-11-10 7:27 ` Christian Zigotzky
2019-11-11 8:12 ` Christian Zigotzky
2019-11-11 8:16 ` Christian Zigotzky
2019-11-11 12:22 ` Christian Zigotzky
2019-11-12 14:41 ` Christoph Hellwig
2019-11-12 22:58 ` Christian Zigotzky
2019-11-13 10:14 ` Christian Zigotzky
2019-11-13 11:02 ` Christoph Hellwig
2018-11-29 17:02 ` use generic DMA mapping code in powerpc V4 Christoph Hellwig
2018-11-30 3:17 ` Benjamin Herrenschmidt
2018-12-08 17:03 ` Christoph Hellwig
2018-12-16 16:51 ` Christoph Hellwig
2018-12-17 1:14 ` Michael Ellerman
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20191105162856.GA15402@lst.de \
--to=hch@lst.de \
--cc=chzigotzky@xenosoft.de \
--cc=iommu@lists.linux-foundation.org \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=olof@lixom.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).